The goal of the proposed work is to understand the pathogenesis of non-cardiac pulmonary edema that occurs after insults elsewhere in the body, e.g., sepsis or non thoracic trauma and shock. Two major hypotheses will be studied. The first is that these kinds of indirect, acute lung injury are mediated via components of blood, i.e., are blood-borne. The second is that many elements of what is commonly called the "acute inflammatory reaction" combine to cause the changes characteristic of acute lung injury. The performance of repeated, limited bronchoalveolar lavage in patients with the adult respiratory distress syndrome (ARDS) is central to the investigation of these hypotheses. Lavage fluid will be collected sequentially and components of the fluid will be studied for proteolytic and chemotactic activity, surface activity , vasoactivity, and neutrophil function. Results of these studies will be correlated with clinical variables collected simultaneously. Mechanisms of lung injury in animals will be investigated concurrently using ethchlorvynol to effect direct, limited pulmonary endothelial injury. More complex inflammatory lung injury will be studied in animals using endotoxin. Potential mediators of these two models, including complement peptides (C5a), proteases, eicosacoids, and oxygen radicals, will be studied separately in animals and in vitro systems. The aims of the animal and in vitro studies are to understand the contribution of formed elements (neutrophils and platelets) and humoral mediators to the endothelial permeability and vasoactive changes of acute lung injury and to elucidate the mechanisms of epithelial injury and repair. Animal models and in vitro systems have been specifically chosen to study aspects that cannot be well-defined in humans, e.g., ultrastructural morphology and biology of limited injury, routes of permeability increase, and mechanisms of repair and structural reorganization.